Background: Colour coded front-of-pack nutrition labelling (‘traffic light labelling’) has been recommended for use in the UK since 2006. The voluntary scheme is used by all the major retailers and some manufacturers. It is not clear how consumers use these labels to make a single decision about the relative healthiness of foods. Our research questions were: Which of the four nutrients on UK traffic light labels (total fat, saturated fat, sugar and salt) has the most influence on decisions? Do green lights or red lights have a greater influence? Are there age and gender differences in how people use the colour and nutrient information?

Methods: We recruited participants from a UK supermarket chain membership list to conduct an online choice experiment in May 2014. We analysed data using multilevel logisitic models with food choices (n = 3321) nested in individuals (n = 187) as the unit of analysis.

Results: A food with more reds was 11.4 (95 % confidence intervals: 10.3, 12.5) times less likely to be chosen as healthy, whereas a food with more greens was 6.1 (5.6, 6.6) times more likely to be chosen as healthy. Foods with better colours on saturated fat and salt were 7.3 (6.7, 8.0) and 7.1 (6.5, 7.8) times more likely to be chosen as healthy – significantly greater than for total fat (odds ratio 4.8 (4.4, 5.3)) and sugar (5.2 (4.7, 5.6)). Results were broadly similar for different genders and age groups.

Conclusions: We found that participants were more concerned with avoiding reds than choosing greens, and that saturated fat and salt had a greater influence on decisions regarding healthiness than total fat and sugar. This could influence decisions about food reformulation and guidance on using nutrition labelling.